It was nearly twenty years from Mendeleev’s first interest in classification to the emergence of his periodic table in 1869. This long pondering and incubation (so similar, in a way, to Darwin’s before he published On the Origin of Species) was perhaps the reason why, when Mendeleev finally published his Principles, he could bring a vastness of knowledge and insight far beyond any of his contemporaries – some of them also had a clear vision of periodicity, but none of them could marshal the overwhelming detail he could.
Mendeleev described how he would write the properties and atomic weights of the elements on cards and ponder and shuffle these constantly on his long railway journeys through Russia, playing a sort of patience or (as he called it) ‘chemical solitaire’, groping for an order, a system that might bring sense to all the elements, their properties and atomic weights.
There was another crucial factor. There had been considerable confusion, for decades, about the atomic weights of many elements. It was only when this was cleared up finally, at the Karlsruhe conference in 1860, that Mendeleev and others could even think of achieving a full taxonomy of the elements. Mendeleev had gone to Karlsruhe with Borodin (this was a musical as well as a chemical journey, for they stopped at many churches en route, trying out the local organs for themselves). With the old, pre-Karlsruhe atomic weights one could get a sense of local triads or groups, but one could not see that there was a numerical relationship between the groups themselves.[42] Only when Cannizzaro showed how reliable atomic weights could be obtained and showed, for example, that the proper atomic weights for the alkaline earth metals (calcium, strontium, and barium) were 40, 88, and 137 (not 20, 44, and 68, as formerly believed) did it become clear how close these were to those of the alkali metals – potassium, rubidium, and cesium. It was this closeness, and in turn the closeness of the atomic weights of the halogens – chlorine, bromine, and iodine – which incited Mendeleev, in 1868, to make a small grid juxtaposing the three groups:
| CI | 35.5 | K | 39 | Ca | 40 |
| Br | 80 | Rb | 85 | Sr | 88 |
| I | 127 | Cs | 133 | Ba | 137 |
And it was at this point, seeing that arranging the three groups of elements in order of atomic weight produced a repetitive pattern – a halogen followed by an alkali metal, followed by an alkaline earth metal – that Mendeleev, feeling this must be a fragment of a larger pattern, leapt to the idea of a periodicity governing all the elements – a Periodic Law.
Mendeleev’s first small table had to be filled in, and then extended in all directions, as if filling up a crossword puzzle; this in itself required some bold speculations. What element, he wondered, was chemically allied with the alkaline earth metals, yet followed lithium in atomic weight? No such element apparently existed – or could it be beryllium, usually considered to be trivalent, with an atomic weight of 14.5? What if it was bivalent instead, with an atomic weight, therefore, not of 14.5 but 9? Then it would follow lithium and fit into the vacant space perfectly.
Moving between conscious calculation and hunch, between intuition and analysis, Mendeleev arrived within a few weeks at a tabulation of thirty-odd elements in order of ascending atomic weight, a tabulation that now suggested there was a recapitulation of properties with every eighth element. And on the night of February 16, 1869, it is said, he had a dream in which he saw almost all of the known elements arrayed in a grand table. The following morning, he committed this to paper.[43]
The logic and pattern of Mendeleev’s table were so clear that certain anomalies stood out at once. Certain elements seemed to be in the wrong places, while certain places had no elements. On the basis of his enormous chemical knowledge, he repositioned half a dozen elements, in defiance of their accepted valency and atomic weights. In doing this, he displayed an audacity that shocked some of his contemporaries (Lothar Meyer, for one, felt it was monstrous to change atomic weights simply because they did not ‘fit’).
In an act of supreme confidence, Mendeleev reserved several empty spaces in his table for elements ‘as yet unknown.’ He asserted that by extrapolating from the properties of the elements above and below (and also, to some extent, from those to either side) one might make a confident prediction as to what these unknown elements would be like. He did exactly this in his 1871 table, predicting in great detail a new element (‘eka-aluminium’) which would come below aluminium in Group III. Four years later just such an element was found, by the French chemist Lecoq de Boisbaudran, and named (either patriotically, or in sly reference to himself, gallus, the cock) gallium.
The exactness of Mendeleev’s prediction was astonishing: he predicted an atomic weight of 68 (Lecoq got 69.9) and a specific gravity of 5.9 (Lecoq got 5.94) and correctly guessed at a great number of gallium’s other physical and chemical properties – its fusibility, its oxides, its salts, its valency. There were some initial discrepancies between Lecoq’s observations and Mendeleev’s predictions, but all of these were rapidly resolved in favor of Mendeleev. Indeed, it was said that Mendeleev had a better grasp of the properties of gallium – an element he had never even seen – than the man who actually discovered it.
Suddenly Mendeleev was no longer seen as a mere speculator or dreamer, but as a man who had discovered a basic law of nature, and now the periodic table was transformed from a pretty but unproven scheme to an invaluable guide which could allow a vast amount of previously unconnected chemical information to be coordinated. It could also be used to suggest all sorts of research in the future, including a systematic search for ‘missing’ elements. ‘Before the promulgation of this law’, Mendeleev was to say nearly twenty years later, ‘chemical elements were mere fragmentary, incidental facts in Nature; there was no special reason to expect the discovery of new elements.’
Now, with Mendeleev’s periodic table, one could not only expect their discovery, but predict their very properties. Mendeleev made two more equally detailed predictions, and these were also confirmed with the discovery of scandium and germanium a few years later.[44] Here, as with gallium, he made his predictions on the basis of analogy and linearity, guessing that the physical and chemical properties of these unknown elements, and their atomic weights, would be between those of the neighboring elements in their vertical groups.[45]
The keystone to the whole table, curiously, was not anticipated by Mendeleev, and perhaps could not have been, for this was not a question of a missing element, but of an entire family or group. When argon was discovered in 1894 – an element which did not seem to fit anywhere in the table – Mendeleev denied at first that it could be an element and thought it was a heavier form of nitrogen (N3, analogous to ozone, 03). But then it became apparent that there was a space for it, right between chlorine and potassium, and indeed, for a whole group coming between the halogens and the alkali metals in every period. This was realized by Lecoq, who went on to predict the atomic weights of the other yet-to-be-discovered gases – and these, indeed, were discovered in short order. With the discovery of helium, neon, krypton, and xenon, it was clear that these gases formed a perfect periodic group, a group so inert, so modest, so unobtrusive, as to have escaped for a century the chemist’s attention.[46] The inert gases were identical in their inability to form compounds; they had a valency, it seemed, of zero.[47]
42
Mendeleev was not the first to see some significance in the atomic weights of elements. When the atomic weights of the alkaline earth metals were established by Berzelius, Dobereiner was struck by the fact that the atomic weight of strontium was just midway between that of calcium and barium. Was this an accident, as Berzelius thought, or an indication of something important and general? Berzelius himself had just discovered selenium in 1817, and at once realized that (in terms of chemical properties) it ‘belonged’ between sulphur and tellurium. Dobereiner went further, and brought out a quantitative relationship too, for its atomic weight was just midway between theirs. And when lithium was discovered later that year (also in Berzelius’s kitchen lab), Dobereiner observed that it completed another triad, of alkali metals: lithium, sodium, and potassium. Feeling, moreover, that the gap in atomic weight between chlorine and iodine was too great, Dobereiner thought (as Davy had before him) that there must be a third element analogous to them, a halogen, with an atomic weight midway between theirs. (This element, bromine, was discovered a few years later.)
There were mixed reactions to Dobereiner’s ‘triads,’ with their implication of a correlation between atomic weight and chemical character. Berzelius and Davy were doubtful of the significance of such ‘numerology,’ as they saw it; but others were intrigued and wondered whether an obscure but fundamental significance was lurking in Dobereiner’s figures.
43
This, at least, is the accepted myth, and one that was later promulgated by Mendeleev himself, somewhat as Kekule was to describe his own discovery of the benzene ring years later, as the result of a dream of snakes biting their own tails. But if one looks at the actual table that Mendeleev sketched, one can see that it is full of transpositions, crossings-out, and calculations in the margins. It shows, in the most graphic way, the creative struggle for understanding which was going on in his mind. Mendeleev did not wake from his dream with all the answers in place, but, more interestingly, perhaps, woke with a sense of revelation, so that within hours he was able to solve many of the questions that had occupied him for years.
44
In an 1889 footnote – even his lectures had footnotes, at least in their printed versions – he added: ‘I foresee some more new elements, but not with the same certitude as before.’ Mendeleev was well aware of the gap between bismuth (with an atomic weight of 209) and thorium (232), and conceived that several elements must exist to fill it. He was most certain of the element immediately following bismuth – ’an element analogous to tellurium, which we may call dvi-tellurium.’ This element, polonium, was discovered by the Curies in 1898, and when finally isolated it had almost all the properties Mendeleev had predicted. (In 1899 Mendeleev visited the Curies in Paris and welcomed radium as his ‘eka-barium.’)
In the final edition of the
He also envisaged, by analogy, some elements following uranium.
45
It is a remarkable example of synchronicity that in the decade following the Karlsruhe conference there emerged not one but
De Chancourtois, a French mineralogist, was the first to devise such a classification, and in 1862 – just eighteen months after Karlsruhe – he inscribed the symbols of twenty-four elements spiraling around a vertical cylinder at heights proportional to their atomic weights, so that elements with similar properties fell one beneath another. Tellurium occupied the midpoint of the helix; hence he called it a ‘telluric screw,’ a
Newlands, in England, was scarcely any luckier. He, too, arranged the known elements by increasing atomic weight, and seeing that every eighth element, apparently, was analogous to the first, he proposed a ‘Law of Octaves,’ saying that ‘the eighth element, starting from a given one, is a kind of repetition of the first, like the 8th note in an octave of music.’ (Had the inert gases been known at the time, it would, of course, have been every ninth element that resembled the first.) A too-literal comparison to music, and the suggestion even that these octaves might be a sort of ‘cosmic music,’ evoked a sarcastic response at the meeting of the Chemical Society at which Newlands presented his theory; it was said that he might have done as well to arrange the elements alphabetically.
There is no doubt that Newlands, even more than de Chancourtois, was very close to a periodic law. Like Mendeleev, Newlands had the courage to invert the order of certain elements when their atomic weight did not match what seemed to be their proper position in his table (though he failed to make any predictions of unknown elements, as Mendeleev did).
Lothar Meyer was also at the Karlsruhe conference and was one of the first to use the revised atomic weights published there in a periodic classification. In 1868 he came up with an elaborate sixteen-columned periodic table (but the publication of this was delayed until after Mendeleev’s table had appeared). Lothar Meyer paid special attention to the physical properties of the elements and their relation to atomic weights, and in 1870 he published a famous graph plotting the atomic weights of the known elements against their ‘atomic volumes’ (this being the ratio of atomic weight to density), a graph that showed high points for the alkali metals and low points for the dense, small-atomed metals of Group VIII (the platinum and iron metals), with all the other elements falling nicely in between. This graph proved a most potent argument for a periodic law and did much to assist the acceptance of Mendeleev’s work.
But at the time of discovering his ‘Natural System,’ Mendeleev was either ignorant of, or denied knowledge of, any attempts comparable to his own. Later, when his name and fame were established, he became more knowledgeable, perhaps more generous, less threatened by the notion of any codiscoverers or forerunners. When, in 1889, he was invited to give the Faraday Lecture in London, he paid a measured tribute to those who had come before him.
46
Cavendish, however, sparking the nitrogen and oxygen of air together, had observed in 1785 that a small amount (‘not more than 1⁄120th part of the whole’) was totally resistant to combination, but no one paid any attention to this until the 1890s.
47
I think I identified at times with the inert gases, and at other times anthropomorphized them, imagining them lonely, cut off, yearning to bond. Was bonding, bonding with other elements, absolutely impossible for them? Might not fluorine, the most active, the most outrageous of the halogens – so eager to combine that it had defeated efforts to isolate it for more than a century – might not fluorine, if given a chance, at least bond with xenon, the heaviest of the inert gases? I pored over tables of physical constants and decided that such a combination was just, in principle, possible.
In the early 1960s, I was overjoyed to hear (even though my mind at this time had moved on to other things) that the American chemist Neil Bartlett had managed to prepare such a compound – a triple compound of platinum, fluorine, and xenon. Xenon fluorides and xenon oxides were subsequently made.
Freeman Dyson has written to me describing his boyhood love of the periodic table and of the inert gases – he, too, saw them, in their bottles, in the Science Museum in South Kensington – and how excited he was years later when he was shown a specimen of barium xenate, seeing the elusive, unreactive gas firmly and beautifully locked up in a crystaclass="underline"
For me too, the periodic table was a passion… As a boy, I stood in front of the display for hours, thinking how wonderful it was that each of these metal foils and jars of gas had its own distinct personality… One of the memorable moments of my life was when Willard Libby came to Princeton with a little jar full of crystals of barium xenate. A stable compound, looking like common salt, but much heavier. This was the magic of chemistry, to see xenon trapped into a crystal.